应用光学, 2015, 36 (2): 220, 网络出版: 2015-04-14   

一种基于特征分析的偏振图像融合方法

Fusion method for polarization images based on analysis of features
作者单位
合肥工业大学 计算机与信息学院, 安徽 合肥 230009
摘要
偏振成像探测能反映出传统光学成像所无法反映的物体的信息,为了克服计算偏振参量图像丢失细节信息的不足,在已有偏振图像融合方法的基础上,提出一种基于综合图像特征的融合方法.相较于已有算法突出图像的某一方面特征,该算法提取图像的灰度特征、纹理特征和形状特征,据此确定融合权值,对图像进行融合,能够较好地反映目标的细节信息,融合后的图像相较于普通光强图像,方差、信息熵分别提高了12.6%、17.5%,平均梯度从0.59提高到1.83.针对该方法用到的特征维数较高的问题,提供了一种简化算法,耗时从30′54″降至13′37″.
Abstract
The polarization imaging detecting can reflect more information that the traditional optics cannot reflect.On the basis of the existing methods of the polarized image fusion,an image fusing method based on comprehensive image features was proposed to overcome the shortcoming that detailed information would be lost during the calculation of polarization parameter images. By extracting the images gray scale,texture and shape features to determine the weight of fusion,the polarized images were fused,and the result of fusion was good,the variance and information entropy increased by 12.6% and 17.5%,respectively,the average gradient increased from 0.59 to 1.83.However,the feature of the method needed higher dimension,a simplified algorithm was provided to improve the operation efficiency which could reduce the time from 30′54″ to 13′37″.

尹海宁, 胡良梅, 范之国. 一种基于特征分析的偏振图像融合方法[J]. 应用光学, 2015, 36(2): 220. Yin Haining, Hu Liangmei, Fan Zhiguo. Fusion method for polarization images based on analysis of features[J]. Journal of Applied Optics, 2015, 36(2): 220.

本文已被 2 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!